This is a live mirror of the Perl 5 development currently hosted at https://github.com/perl/perl5
https://perl5.git.perl.org / perl5.git / blame
? search:
summary | shortlog | log | commit | commitdiff | tree
blob | blame (incremental) | history | HEAD
utf8.c: Nits in pod
[perl5.git] / utf8.h
CommitLineData
a0ed51b3
LW		 1/* utf8.h
2 *
83706693 3 * Copyright (C) 2000, 2001, 2002, 2005, 2006, 2007, 2009 by Larry Wall and others
a0ed51b3
LW		 4 *
5 * You may distribute under the terms of either the GNU General Public
6 * License or the Artistic License, as specified in the README file.
7 *
8 */
9
39e02b42 10/* Use UTF-8 as the default script encoding?
1e54db1a 11 * Turning this on will break scripts having non-UTF-8 binary
39e02b42
JH		 12 * data (such as Latin-1) in string literals. */
13#ifdef USE_UTF8_SCRIPTS
14# define USE_UTF8_IN_NAMES (!IN_BYTES)
15#else
16# define USE_UTF8_IN_NAMES (PL_hints & HINT_UTF8)
17#endif
18
fd7cb289
RGS		 19/* Source backward compatibility. */
20#define uvuni_to_utf8(d, uv) uvuni_to_utf8_flags(d, uv, 0)
21#define is_utf8_string_loc(s, len, ep) is_utf8_string_loclen(s, len, ep, 0)
22
e6226b18
KW		 23/*
24=for apidoc ibcmp_utf8
25
26This is a synonym for (! foldEQ_utf8())
27
28=cut
29*/
30#define ibcmp_utf8(s1, pe1, l1, u1, s2, pe2, l2, u2) \
31 cBOOL(! foldEQ_utf8(s1, pe1, l1, u1, s2, pe2, l2, u2))
32
1d72bdf6
NIS		 33#ifdef EBCDIC
34/* The equivalent of these macros but implementing UTF-EBCDIC
35 are in the following header file:
36 */
37
38#include "utfebcdic.h"
fd7cb289 39
d06134e5 40#else /* ! EBCDIC */
73c4f7a1
GS		 41START_EXTERN_C
42
a0ed51b3 43#ifdef DOINIT
6f06b55f 44EXTCONST unsigned char PL_utf8skip[] = {
a0ed51b3
LW		 451,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1, /* ascii */
461,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1, /* ascii */
471,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1, /* ascii */
481,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1, /* ascii */
491,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1, /* bogus */
501,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1, /* bogus */
512,2,2,2,2,2,2,2,2,2,2,2,2,2,2,2,2,2,2,2,2,2,2,2,2,2,2,2,2,2,2,2, /* scripts */
3c77ea2b
GS		 523,3,3,3,3,3,3,3,3,3,3,3,3,3,3,3,4,4,4,4,4,4,4,4,5,5,5,5,6,6, /* cjk etc. */
537,13, /* Perl extended (not UTF-8). Up to 72bit allowed (64-bit + reserved). */
a0ed51b3
LW		 54};
55#else
6f06b55f 56EXTCONST unsigned char PL_utf8skip[];
a0ed51b3
LW		 57#endif
58
73c4f7a1 59END_EXTERN_C
7e2040f0 60
1d72bdf6
NIS		 61/* Native character to iso-8859-1 */
62#define NATIVE_TO_ASCII(ch) (ch)
63#define ASCII_TO_NATIVE(ch) (ch)
64/* Transform after encoding */
65#define NATIVE_TO_UTF(ch) (ch)
66#define UTF_TO_NATIVE(ch) (ch)
67/* Transforms in wide UV chars */
68#define UNI_TO_NATIVE(ch) (ch)
69#define NATIVE_TO_UNI(ch) (ch)
70/* Transforms in invariant space */
71#define NATIVE_TO_NEED(enc,ch) (ch)
72#define ASCII_TO_NEED(enc,ch) (ch)
d7578b48 73
d06134e5 74/* As there are no translations, avoid the function wrapper */
1754c1a1
NC		 75#define utf8n_to_uvchr utf8n_to_uvuni
76#define uvchr_to_utf8 uvuni_to_utf8
2b9d42f0 77
877d9f0d 78/*
9041c2e3 79
8c007b5a 80 The following table is from Unicode 3.2.
877d9f0d
JH		 81
82 Code Points 1st Byte 2nd Byte 3rd Byte 4th Byte
83
375122d7 84 U+0000..U+007F 00..7F
e1b711da 85 U+0080..U+07FF * C2..DF 80..BF
37e2e78e 86 U+0800..U+0FFF E0 * A0..BF 80..BF
375122d7 87 U+1000..U+CFFF E1..EC 80..BF 80..BF
e1b711da 88 U+D000..U+D7FF ED 80..9F 80..BF
37e2e78e 89 U+D800..U+DFFF +++++++ utf16 surrogates, not legal utf8 +++++++
375122d7 90 U+E000..U+FFFF EE..EF 80..BF 80..BF
37e2e78e 91 U+10000..U+3FFFF F0 * 90..BF 80..BF 80..BF
877d9f0d
JH		 92 U+40000..U+FFFFF F1..F3 80..BF 80..BF 80..BF
93 U+100000..U+10FFFF F4 80..8F 80..BF 80..BF
94
e1b711da 95Note the gaps before several of the byte entries above marked by '*'. These are
37e2e78e
KW		 96caused by legal UTF-8 avoiding non-shortest encodings: it is technically
97possible to UTF-8-encode a single code point in different ways, but that is
98explicitly forbidden, and the shortest possible encoding should always be used
99(and that is what Perl does).
8c007b5a 100
877d9f0d
JH		 101 */
102
8c007b5a
JH		 103/*
104 Another way to look at it, as bits:
105
106 Code Points 1st Byte 2nd Byte 3rd Byte 4th Byte
107
108 0aaaaaaa 0aaaaaaa
109 00000bbbbbaaaaaa 110bbbbb 10aaaaaa
110 ccccbbbbbbaaaaaa 1110cccc 10bbbbbb 10aaaaaa
111 00000dddccccccbbbbbbaaaaaa 11110ddd 10cccccc 10bbbbbb 10aaaaaa
112
113As you can see, the continuation bytes all begin with C<10>, and the
e1b711da 114leading bits of the start byte tell how many bytes there are in the
8c007b5a
JH		 115encoded character.
116
65ab9279
TC		 117Perl's extended UTF-8 means we can have start bytes up to FF.
118
8c007b5a
JH		 119*/
120
c4d5f83a 121#define UNI_IS_INVARIANT(c) (((UV)c) < 0x80)
d06134e5
KW		 122/* Note that C0 and C1 are invalid in legal UTF8, so the lower bound of the
123 * below might ought to be C2 */
65ab9279 124#define UTF8_IS_START(c) (((U8)c) >= 0xc0)
c512ce4f
JH		 125#define UTF8_IS_CONTINUATION(c) (((U8)c) >= 0x80 && (((U8)c) <= 0xbf))
126#define UTF8_IS_CONTINUED(c) (((U8)c) & 0x80)
db42d148 127#define UTF8_IS_DOWNGRADEABLE_START(c) (((U8)c & 0xfc) == 0xc0)
8850bf83 128
22901f30
RGS		 129#define UTF_START_MARK(len) (((len) > 7) ? 0xFF : (0xFE << (7-(len))))
130#define UTF_START_MASK(len) (((len) >= 7) ? 0x00 : (0x1F >> ((len)-2)))
1d72bdf6
NIS		 131
132#define UTF_CONTINUATION_MARK 0x80
133#define UTF_ACCUMULATION_SHIFT 6
134#define UTF_CONTINUATION_MASK ((U8)0x3f)
c512ce4f 135
1d68d6cd 136#ifdef HAS_QUAD
5bbb0b5a 137#define UNISKIP(uv) ( (uv) < 0x80 ? 1 : \
1d68d6cd
SC		 138 (uv) < 0x800 ? 2 : \
139 (uv) < 0x10000 ? 3 : \
140 (uv) < 0x200000 ? 4 : \
141 (uv) < 0x4000000 ? 5 : \
142 (uv) < 0x80000000 ? 6 : \
9041c2e3 143 (uv) < UTF8_QUAD_MAX ? 7 : 13 )
1d68d6cd
SC		 144#else
145/* No, I'm not even going to *TRY* putting #ifdef inside a #define */
5bbb0b5a 146#define UNISKIP(uv) ( (uv) < 0x80 ? 1 : \
1d68d6cd
SC		 147 (uv) < 0x800 ? 2 : \
148 (uv) < 0x10000 ? 3 : \
149 (uv) < 0x200000 ? 4 : \
150 (uv) < 0x4000000 ? 5 : \
151 (uv) < 0x80000000 ? 6 : 7 )
152#endif
153
d06134e5
KW		 154#endif /* EBCDIC vs ASCII */
155
156/* Rest of these are attributes of Unicode and perl's internals rather than the
157 * encoding, or happen to be the same in both ASCII and EBCDIC (at least at
158 * this level; the macros that some of these call may have different
159 * definitions in the two encodings */
160
161#define NATIVE8_TO_UNI(ch) NATIVE_TO_ASCII(ch) /* a clearer synonym */
162
163#define UTF8_ACCUMULATE(old, new) (((old) << UTF_ACCUMULATION_SHIFT) | (((U8)new) & UTF_CONTINUATION_MASK))
164
2950f2a7
KW		 165/* Convert a two (not one) byte utf8 character to a unicode code point value.
166 * Needs just one iteration of accumulate. Should not be used unless it is
167 * known that the two bytes are legal: 1) two-byte start, and 2) continuation.
168 * Note that the result can be larger than 255 if the input character is not
169 * downgradable */
170#define TWO_BYTE_UTF8_TO_UNI(HI, LO) \
171 UTF8_ACCUMULATE((NATIVE_TO_UTF(HI) & UTF_START_MASK(2)), \
172 NATIVE_TO_UTF(LO))
173
d06134e5
KW		 174#define UTF8SKIP(s) PL_utf8skip[*(const U8*)(s)]
175
176#define UTF8_IS_INVARIANT(c) UNI_IS_INVARIANT(NATIVE_TO_UTF(c))
177#define NATIVE_IS_INVARIANT(c) UNI_IS_INVARIANT(NATIVE8_TO_UNI(c))
178
179#define MAX_PORTABLE_UTF8_TWO_BYTE 0x3FF /* constrained by EBCDIC */
180
181/* The macros in the next sets are used to generate the two utf8 or utfebcdic
182 * bytes from an ordinal that is known to fit into two bytes; it must be less
183 * than 0x3FF to work across both encodings. */
184/* Nocast allows these to be used in the case label of a switch statement */
0072151e 185#define UTF8_TWO_BYTE_HI_nocast(c) UTF_TO_NATIVE(((c) >> UTF_ACCUMULATION_SHIFT) | (0xFF & UTF_START_MARK(2)))
14c28784 186#define UTF8_TWO_BYTE_LO_nocast(c) UTF_TO_NATIVE(((c) & UTF_CONTINUATION_MASK) | UTF_CONTINUATION_MARK)
d06134e5
KW		 187
188#define UTF8_TWO_BYTE_HI(c) ((U8) (UTF8_TWO_BYTE_HI_nocast(c)))
189#define UTF8_TWO_BYTE_LO(c) ((U8) (UTF8_TWO_BYTE_LO_nocast(c)))
190
191/* This name is used when the source is a single byte */
192#define UTF8_EIGHT_BIT_HI(c) UTF8_TWO_BYTE_HI((U8)(c))
193#define UTF8_EIGHT_BIT_LO(c) UTF8_TWO_BYTE_LO((U8)(c))
194
7e2040f0
GS		 195/*
196 * Note: we try to be careful never to call the isXXX_utf8() functions
d06134e5
KW		 197 * unless we're pretty sure we've seen the beginning of a UTF-8 or UTFEBCDIC
198 * character. Otherwise we risk loading in the heavy-duty swash_init and
199 * swash_fetch routines unnecessarily.
7e2040f0 200 */
d06134e5 201#define isIDFIRST_lazy_if(p,c) ((IN_BYTES || (!c || ! UTF8_IS_START(*((const U8*)p)))) \
7e2040f0 202 ? isIDFIRST(*(p)) \
e1ec3a88 203 : isIDFIRST_utf8((const U8*)p))
d06134e5 204#define isALNUM_lazy_if(p,c) ((IN_BYTES || (!c || ! UTF8_IS_START(*((const U8*)p)))) \
7e2040f0 205 ? isALNUM(*(p)) \
e1ec3a88 206 : isALNUM_utf8((const U8*)p))
1d72bdf6 207
7e2040f0
GS		 208#define isIDFIRST_lazy(p) isIDFIRST_lazy_if(p,1)
209#define isALNUM_lazy(p) isALNUM_lazy_if(p,1)
3bd709b1 210
89ebb4a3
JH		 211#define UTF8_MAXBYTES 13
212/* How wide can a single UTF-8 encoded character become in bytes.
213 * NOTE: Strictly speaking Perl's UTF-8 should not be called UTF-8
214 * since UTF-8 is an encoding of Unicode and given Unicode's current
215 * upper limit only four bytes is possible. Perl thinks of UTF-8
216 * as a way to encode non-negative integers in a binary format. */
217#define UTF8_MAXLEN UTF8_MAXBYTES
218
89ebb4a3
JH		 219/* The maximum number of UTF-8 bytes a single Unicode character can
220 * uppercase/lowercase/fold into; this number depends on the Unicode
221 * version. An example of maximal expansion is the U+03B0 which
222 * uppercases to U+03C5 U+0308 U+0301. The Unicode databases that
d06134e5 223 * tell these things are UnicodeData.txt, CaseFolding.txt, and
89ebb4a3
JH		 224 * SpecialCasing.txt. */
225#define UTF8_MAXBYTES_CASE 6
3bd709b1 226
a98fe34d 227#define IN_BYTES (CopHINTS_get(PL_curcop) & HINT_BYTES)
0064a8a9 228#define DO_UTF8(sv) (SvUTF8(sv) && !IN_BYTES)
1863b879 229#define IN_UNI_8_BIT ( (CopHINTS_get(PL_curcop) & HINT_UNI_8_BIT) \
61fc5122 230 && ! IN_LOCALE_RUNTIME && ! IN_BYTES)
1d72bdf6 231
c76687c5
KW		 232#define UTF8_ALLOW_EMPTY 0x0001 /* Allow a zero length string */
233
234/* Allow first byte to be a continuation byte */
1d72bdf6 235#define UTF8_ALLOW_CONTINUATION 0x0002
c76687c5
KW		 236
237/* Allow second... bytes to be non-continuation bytes */
1d72bdf6 238#define UTF8_ALLOW_NON_CONTINUATION 0x0004
37e2e78e 239#define UTF8_ALLOW_FE_FF 0x0008 /* Allow FE or FF start bytes, \
ff97e5cf 240 yields above 0x7fffFFFF = 31 bits */
37e2e78e 241#define UTF8_ALLOW_SHORT 0x0010 /* expecting more bytes */
1d72bdf6 242#define UTF8_ALLOW_SURROGATE 0x0020
9f7f3913 243#define UTF8_ALLOW_FFFF 0x0040 /* Allow UNICODE_ILLEGAL */
37e2e78e 244#define UTF8_ALLOW_LONG 0x0080 /* expecting fewer bytes */
1d72bdf6 245#define UTF8_ALLOW_ANYUV (UTF8_ALLOW_EMPTY|UTF8_ALLOW_FE_FF|\
61843245 246 UTF8_ALLOW_SURROGATE|UTF8_ALLOW_FFFF)
c867b360 247#define UTF8_ALLOW_ANY 0x00FF
e83d50c9 248#define UTF8_CHECK_ONLY 0x0200
9f7f3913
TS		 249#define UTF8_ALLOW_DEFAULT (ckWARN(WARN_UTF8) ? 0 : \
250 UTF8_ALLOW_ANYUV)
1d72bdf6 251
7131f24d
KW		 252/* Surrogates, non-character code points and above-Unicode code points are
253 * problematic in some contexts. This allows code that needs to check for
254 * those to to quickly exclude the vast majority of code points it will
255 * encounter */
256#ifdef EBCDIC
257# define UTF8_FIRST_PROBLEMATIC_CODE_POINT_FIRST_BYTE UTF_TO_NATIVE(0xF1)
258#else
259# define UTF8_FIRST_PROBLEMATIC_CODE_POINT_FIRST_BYTE 0xED
260#endif
261
262/* ASCII EBCDIC I8
263 * U+D7FF: \xED\x9F\xBF \xF1\xB5\xBF\xBF last before surrogates
264 * U+D800: \xED\xA0\x80 \xF1\xB6\xA0\xA0 1st surrogate
265 * U+DFFF: \xED\xBF\xBF \xF1\xB7\xBF\xBF final surrogate
266 * U+E000: \xEE\x80\x80 \xF1\xB8\xA0\xA0 next after surrogates
267 */
268#ifdef EBCDIC /* Both versions assume well-formed UTF8 */
269# define UTF8_IS_SURROGATE(s) (*(s) == UTF_TO_NATIVE(0xF1) \
270 && (*((s) +1) == UTF_TO_NATIVE(0xB6)) || *((s) + 1) == UTF_TO_NATIVE(0xB7))
271#else
272# define UTF8_IS_SURROGATE(s) (*(s) == 0xED && *((s) + 1) >= 0xA0)
273#endif
274
275/* ASCII EBCDIC I8
276 * U+10FFFF: \xF4\x8F\xBF\xBF \xF9\xA1\xBF\xBF\xBF max legal Unicode
277 * U+110000: \xF4\x90\x80\x80 \xF9\xA2\xA0\xA0\xA0
278 * U+110001: \xF4\x90\x80\x81 \xF9\xA2\xA0\xA0\xA1
279 */
280#ifdef EBCDIC /* Both versions assume well-formed UTF8 */
281# define UTF8_IS_SUPER(s) (*(s) >= UTF_TO_NATIVE(0xF9) \
282 && (*(s) > UTF_TO_NATIVE(0xF9) || (*((s) + 1) >= UTF_TO_NATIVE(0xA2))))
283#else
284# define UTF8_IS_SUPER(s) (*(s) >= 0xF4 \
285 && (*(s) > 0xF4 || (*((s) + 1) >= 0x90)))
286#endif
287
288/* ASCII EBCDIC I8
289 * U+FDCF: \xEF\xB7\x8F \xF1\xBF\xAE\xAF last before non-char block
290 * U+FDD0: \xEF\xB7\x90 \xF1\xBF\xAE\xB0 first non-char in block
291 * U+FDEF: \xEF\xB7\xAF \xF1\xBF\xAF\xAF last non-char in block
292 * U+FDF0: \xEF\xB7\xB0 \xF1\xBF\xAF\xB0 first after non-char block
293 * U+FFFF: \xEF\xBF\xBF \xF1\xBF\xBF\xBF
294 * U+1FFFF: \xF0\x9F\xBF\xBF \xF3\xBF\xBF\xBF
295 * U+2FFFF: \xF0\xAF\xBF\xBF \xF5\xBF\xBF\xBF
296 * U+3FFFF: \xF0\xBF\xBF\xBF \xF7\xBF\xBF\xBF
297 * U+4FFFF: \xF1\x8F\xBF\xBF \xF8\xA9\xBF\xBF\xBF
298 * U+5FFFF: \xF1\x9F\xBF\xBF \xF8\xAB\xBF\xBF\xBF
299 * U+6FFFF: \xF1\xAF\xBF\xBF \xF8\xAD\xBF\xBF\xBF
300 * U+7FFFF: \xF1\xBF\xBF\xBF \xF8\xAF\xBF\xBF\xBF
301 * U+8FFFF: \xF2\x8F\xBF\xBF \xF8\xB1\xBF\xBF\xBF
302 * U+9FFFF: \xF2\x9F\xBF\xBF \xF8\xB3\xBF\xBF\xBF
303 * U+AFFFF: \xF2\xAF\xBF\xBF \xF8\xB5\xBF\xBF\xBF
304 * U+BFFFF: \xF2\xBF\xBF\xBF \xF8\xB7\xBF\xBF\xBF
305 * U+CFFFF: \xF3\x8F\xBF\xBF \xF8\xB9\xBF\xBF\xBF
306 * U+DFFFF: \xF3\x9F\xBF\xBF \xF8\xBB\xBF\xBF\xBF
307 * U+EFFFF: \xF3\xAF\xBF\xBF \xF8\xBD\xBF\xBF\xBF
308 * U+FFFFF: \xF3\xBF\xBF\xBF \xF8\xBF\xBF\xBF\xBF
309 * U+10FFFF: \xF4\x8F\xBF\xBF \xF9\xA1\xBF\xBF\xBF
310 */
311#define UTF8_IS_NONCHAR_(s) ( \
312 *(s) >= UTF8_FIRST_PROBLEMATIC_CODE_POINT_FIRST_BYTE \
313 && ! UTF8_IS_SUPER(s) \
314 && UTF8_IS_NONCHAR_GIVEN_THAT_NON_SUPER_AND_GE_FIRST_PROBLEMATIC(s) \
315
316#ifdef EBCDIC /* Both versions assume well-formed UTF8 */
317# define UTF8_IS_NONCHAR_GIVEN_THAT_NON_SUPER_AND_GE_PROBLEMATIC(s) \
318 ((*(s) == UTF_TO_NATIVE(0xF1) \
319 && (*((s) + 1) == UTF_TO_NATIVE(0xBF) \
320 && ((*((s) + 2) == UTF_TO_NATIVE(0xAE) \
321 && *((s) + 3) >= UTF_TO_NATIVE(0xB0)) \
322 || (*((s) + 2) == UTF_TO_NATIVE(0xAF) \
323 && *((s) + 3) <= UTF_TO_NATIVE(0xAF))))) \
324 || (UTF8SKIP(*(s)) > 3 \
325 /* (These were all derived by inspection and experimentation with an */ \
326 /* editor) The next line checks the next to final byte in the char */ \
327 && *((s) + UTF8SKIP(*(s)) - 2) == UTF_TO_NATIVE(0xBF) \
328 && *((s) + UTF8SKIP(*(s)) - 3) == UTF_TO_NATIVE(0xBF) \
329 && (NATIVE_TO_UTF(*((s) + UTF8SKIP(*(s)) - 4)) & 0x81) == 0x81 \
330 && (NATIVE_TO_UTF(*((s) + UTF8SKIP(*(s)) - 1)) & 0xBE) == 0XBE))
331#else
332# define UTF8_IS_NONCHAR_GIVEN_THAT_NON_SUPER_AND_GE_PROBLEMATIC(s) \
333 ((*(s) == 0xEF \
334 && ((*((s) + 1) == 0xB7 && (*((s) + 2) >= 0x90 && (*((s) + 2) <= 0xAF)))\
335 /* Gets U+FFF[EF] */ \
336 || (*((s) + 1) == 0xBF && ((*((s) + 2) & 0xBE) == 0xBE)))) \
337 || ((*((s) + 2) == 0xBF \
338 && (*((s) + 3) & 0xBE) == 0xBE \
339 /* Excludes things like U+10FFE = \xF0\x90\xBF\xBE */ \
340 && (*((s) + 1) & 0x8F) == 0x8F)))
341#endif
342
c867b360
JH		 343#define UNICODE_SURROGATE_FIRST 0xD800
344#define UNICODE_SURROGATE_LAST 0xDFFF
345#define UNICODE_REPLACEMENT 0xFFFD
346#define UNICODE_BYTE_ORDER_MARK 0xFEFF
347#define UNICODE_ILLEGAL 0xFFFF
1d72bdf6 348
b851fbc1 349/* Though our UTF-8 encoding can go beyond this,
c76687c5 350 * let's be conservative and do as Unicode says. */
b851fbc1
JH		 351#define PERL_UNICODE_MAX 0x10FFFF
352
353#define UNICODE_ALLOW_SURROGATE 0x0001 /* Allow UTF-16 surrogates (EVIL) */
354#define UNICODE_ALLOW_FDD0 0x0002 /* Allow the U+FDD0...U+FDEF */
9f7f3913
TS		 355#define UNICODE_ALLOW_FFFF 0x0004 /* Allow U+FFF[EF], U+1FFF[EF], ... */
356#define UNICODE_ALLOW_SUPER 0x0008 /* Allow past 0x10FFFF */
c867b360 357#define UNICODE_ALLOW_ANY 0x000F
b851fbc1 358
1d72bdf6
NIS		 359#define UNICODE_IS_SURROGATE(c) ((c) >= UNICODE_SURROGATE_FIRST && \
360 (c) <= UNICODE_SURROGATE_LAST)
a10ec373 361#define UNICODE_IS_REPLACEMENT(c) ((c) == UNICODE_REPLACEMENT)
872c91ae 362#define UNICODE_IS_BYTE_ORDER_MARK(c) ((c) == UNICODE_BYTE_ORDER_MARK)
1d72bdf6 363#define UNICODE_IS_ILLEGAL(c) ((c) == UNICODE_ILLEGAL)
7131f24d
KW		 364#define UNICODE_IS_NONCHAR(c) ((c >= 0xFDD0 && c <= 0xFDEF) \
365 /* The other noncharacters end in FFFE or FFFF, which \
366 * the mask below catches both of, but beyond the last \
367 * official unicode code point, they aren't \
368 * noncharacters, since those aren't Unicode \
369 * characters at all */ \
370 || ((((c & 0xFFFE) == 0xFFFE)) && ! UNICODE_IS_SUPER(c)))
371#define UNICODE_IS_SUPER(c) ((c) > PERL_UNICODE_MAX)
372#define UNICODE_IS_FE_FF(c) ((c) > 0x7FFFFFFF)
1d72bdf6 373
f067b878
NA		 374#ifdef HAS_QUAD
375# define UTF8_QUAD_MAX UINT64_C(0x1000000000)
376#endif
3bd709b1 377
09091399
JH		 378#define UNICODE_GREEK_CAPITAL_LETTER_SIGMA 0x03A3
379#define UNICODE_GREEK_SMALL_LETTER_FINAL_SIGMA 0x03C2
380#define UNICODE_GREEK_SMALL_LETTER_SIGMA 0x03C3
9dcbe121 381#define GREEK_SMALL_LETTER_MU 0x03BC
09091399 382
9e55ce06 383#define UNI_DISPLAY_ISPRINT 0x0001
c728cb41
JH		 384#define UNI_DISPLAY_BACKSLASH 0x0002
385#define UNI_DISPLAY_QQ (UNI_DISPLAY_ISPRINT|UNI_DISPLAY_BACKSLASH)
386#define UNI_DISPLAY_REGEX (UNI_DISPLAY_ISPRINT|UNI_DISPLAY_BACKSLASH)
9e55ce06 387
5cd46e1f
KW		 388#ifndef EBCDIC
389# define LATIN_SMALL_LETTER_SHARP_S 0x00DF
390# define LATIN_SMALL_LETTER_Y_WITH_DIAERESIS 0x00FF
391# define MICRO_SIGN 0x00B5
392#endif
393
394#define ANYOF_FOLD_SHARP_S(node, input, end) \
395 (ANYOF_BITMAP_TEST(node, LATIN_SMALL_LETTER_SHARP_S) && \
3ff7ceb3 396 (ANYOF_FLAGS(node) & ANYOF_NONBITMAP) && \
39065660 397 (ANYOF_FLAGS(node) & ANYOF_LOC_NONBITMAP_FOLD) && \
07b6858f
JH		 398 ((end) > (input) + 1) && \
399 toLOWER((input)[0]) == 's' && \
400 toLOWER((input)[1]) == 's')
ebc501f0 401#define SHARP_S_SKIP 2
3b0fc154 402
77263263
TS		 403#ifdef EBCDIC
404/* IS_UTF8_CHAR() is not ported to EBCDIC */
405#else
3b0fc154
JH		 406#define IS_UTF8_CHAR_1(p) \
407 ((p)[0] <= 0x7F)
408#define IS_UTF8_CHAR_2(p) \
409 ((p)[0] >= 0xC2 && (p)[0] <= 0xDF && \
410 (p)[1] >= 0x80 && (p)[1] <= 0xBF)
411#define IS_UTF8_CHAR_3a(p) \
412 ((p)[0] == 0xE0 && \
413 (p)[1] >= 0xA0 && (p)[1] <= 0xBF && \
414 (p)[2] >= 0x80 && (p)[2] <= 0xBF)
415#define IS_UTF8_CHAR_3b(p) \
416 ((p)[0] >= 0xE1 && (p)[0] <= 0xEC && \
417 (p)[1] >= 0x80 && (p)[1] <= 0xBF && \
418 (p)[2] >= 0x80 && (p)[2] <= 0xBF)
419#define IS_UTF8_CHAR_3c(p) \
420 ((p)[0] == 0xED && \
421 (p)[1] >= 0x80 && (p)[1] <= 0xBF && \
422 (p)[2] >= 0x80 && (p)[2] <= 0xBF)
423/* In IS_UTF8_CHAR_3c(p) one could use
424 * (p)[1] >= 0x80 && (p)[1] <= 0x9F
425 * if one wanted to exclude surrogates. */
426#define IS_UTF8_CHAR_3d(p) \
427 ((p)[0] >= 0xEE && (p)[0] <= 0xEF && \
428 (p)[1] >= 0x80 && (p)[1] <= 0xBF && \
429 (p)[2] >= 0x80 && (p)[2] <= 0xBF)
430#define IS_UTF8_CHAR_4a(p) \
431 ((p)[0] == 0xF0 && \
432 (p)[1] >= 0x90 && (p)[1] <= 0xBF && \
433 (p)[2] >= 0x80 && (p)[2] <= 0xBF && \
434 (p)[3] >= 0x80 && (p)[3] <= 0xBF)
435#define IS_UTF8_CHAR_4b(p) \
436 ((p)[0] >= 0xF1 && (p)[0] <= 0xF3 && \
437 (p)[1] >= 0x80 && (p)[1] <= 0xBF && \
438 (p)[2] >= 0x80 && (p)[2] <= 0xBF && \
439 (p)[3] >= 0x80 && (p)[3] <= 0xBF)
440/* In IS_UTF8_CHAR_4c(p) one could use
441 * (p)[0] == 0xF4
442 * if one wanted to stop at the Unicode limit U+10FFFF.
443 * The 0xF7 allows us to go to 0x1fffff (0x200000 would
444 * require five bytes). Not doing any further code points
445 * since that is not needed (and that would not be strict
446 * UTF-8, anyway). The "slow path" in Perl_is_utf8_char()
447 * will take care of the "extended UTF-8". */
448#define IS_UTF8_CHAR_4c(p) \
a18d6e6e 449 ((p)[0] >= 0xF4 && (p)[0] <= 0xF7 && \
3b0fc154
JH		 450 (p)[1] >= 0x80 && (p)[1] <= 0xBF && \
451 (p)[2] >= 0x80 && (p)[2] <= 0xBF && \
452 (p)[3] >= 0x80 && (p)[3] <= 0xBF)
453
454#define IS_UTF8_CHAR_3(p) \
455 (IS_UTF8_CHAR_3a(p) || \
456 IS_UTF8_CHAR_3b(p) || \
457 IS_UTF8_CHAR_3c(p) || \
458 IS_UTF8_CHAR_3d(p))
459#define IS_UTF8_CHAR_4(p) \
460 (IS_UTF8_CHAR_4a(p) || \
461 IS_UTF8_CHAR_4b(p) || \
462 IS_UTF8_CHAR_4c(p))
463
464/* IS_UTF8_CHAR(p) is strictly speaking wrong (not UTF-8) because it
465 * (1) allows UTF-8 encoded UTF-16 surrogates
466 * (2) it allows code points past U+10FFFF.
467 * The Perl_is_utf8_char() full "slow" code will handle the Perl
468 * "extended UTF-8". */
469#define IS_UTF8_CHAR(p, n) \
470 ((n) == 1 ? IS_UTF8_CHAR_1(p) : \
471 (n) == 2 ? IS_UTF8_CHAR_2(p) : \
472 (n) == 3 ? IS_UTF8_CHAR_3(p) : \
473 (n) == 4 ? IS_UTF8_CHAR_4(p) : 0)
474
768c67ee
JH		 475#define IS_UTF8_CHAR_FAST(n) ((n) <= 4)
476
77263263 477#endif /* IS_UTF8_CHAR() for UTF-8 */
e9a8c099
MHM		 478
479/*
480 * Local variables:
481 * c-indentation-style: bsd
482 * c-basic-offset: 4
483 * indent-tabs-mode: t
484 * End:
485 *
486 * ex: set ts=8 sts=4 sw=4 noet:
487 */
Unnamed repository; edit this file 'description' to name the repository.